Quality Assurance in Clinical Data Management

Abstract

Clinical data management is involved in ensuring data quality and integrity from the time a draft of a clinical protocol is issued through the final statistical analysis of the resultant data. This involves careful consideration of the clinical forms and schedule of patient visits; feedback to clinical monitors of systematic errors by investigators; feedback to data processing on data entry errors, etc. These actions involve medical staff, data processing staff, and biostatisticians. Because these units are independent it has been our experience that a quality assurance unit dedicated to data quality assurance and integrity would be useful. The purpose of this unit is to:

1. 1.

Identify data processing errors, correct such errors, and provide feedback to data managers and medical staff during and after completion of a study.

2.

Identify and document protocol violations, missing or unprocessed documents, early terminations, and any other special situations regarding a patient that a biostatistician needs to know.

3.

Report any special data processing situations or deviations from coding conventions.

4.

Identify and provide feedback concerning problems with the clinical data base.

This paper discusses the organization, procedures, and technical methods used by this unit to certify data processing quality and data integrity.